Forum Discussion
3 Replies
- LNATIONDEVMHCP Partner
I've had a play. Some feedback:
For the 3d model the options you get when you search something simple like "A wooden climbing frame" are pretty terrible, it does however do a clever job of rendering the image to the 3d model. It would be neat if you could upload your own picture or draw something in 2d and then have it rendered into 3d.
I used the sky and texture generator in my first test maze project, they were straightforward to use. It took me a few times to get a texture I was relatively happy with it and looked better on mobile than in VR but I suspect that is just my poor implementation.
The sound tools could be better, my first use case was the need for a human countdown (3,2,1,go) it can't do this though. Perhaps you should think about adding text to speech functionality. It gave me weird sound effects which I guess is what the tool is supposed to be used for. The ambient background could be improved also, I ended up using an external tool to generate my background noise.
I haven't tried the script generation tools but this should be baked into visual studios via a plugin or something. Currently as I hope you're aware co-pilot is not very well trained with your APIs and struggles to implement basic functionality. I have actually disabled it while I work on your scripts.
- KevinJardineStart Partner
Very exciting. I was hoping to use this to generate some furniture and other structures for the star stations in my Galactic Treasure Hunt game. However, I found that the models generated were more detailed than I wanted and the corresponding textures not detailed enough (or at least too fuzzy).
I will keep experimenting but for now I think low poly generation and texturing in Blender will continue to give me closer to what I need.
- KevinJardineStart Partner
Providing more details on the texture seems to help. This model was generated using "A cartoony red green and blue solid colored low poly retro robot".